It doesn’t matter whether or not an individual’s precise voice is utilized in an imitation or not, Rothman says, solely whether or not that audio confuses listeners. Within the authorized system, there’s a massive distinction between imitation and easily recording one thing “within the model” of another person. “Nobody owns a method,” she says.
Different authorized specialists don’t see what OpenAI did as a clear-cut impersonation. “I believe that any potential ‘proper of publicity’ declare from Scarlett Johansson in opposition to OpenAI can be pretty weak given the one superficial similarity between the ‘Sky’ actress’ voice and Johansson, underneath the related case legislation,” Colorado legislation professor Harry Surden wrote on X on Tuesday. Frye, too, has doubts. “OpenAI didn’t say and even suggest it was providing the true Scarlett Johansson, solely a simulation. If it used her identify or picture to promote its product, that may be a right-of-publicity drawback. However merely cloning the sound of her voice in all probability isn’t,” he says.
However that doesn’t imply OpenAI is essentially within the clear. “Juries are unpredictable,” Surden added.
Frye can also be unsure how any case would possibly play out, as a result of he says proper of publicity is a reasonably “esoteric” space of legislation. There aren’t any federal right-of-publicity legal guidelines in the USA, solely a patchwork of state statutes. “It’s a large number,” he says, though Johansson may convey a swimsuit in California, which has pretty sturdy right-of-publicity legal guidelines.
OpenAI’s probabilities of defending a right-of-publicity swimsuit may very well be weakened by a one-word publish on X—“her”—from Sam Altman on the day of final week’s demo. It was broadly interpreted as a reference to Her and Johansson’s efficiency. “It appears like AI from the flicks,” Altman wrote in a weblog publish that day.
To Grimmelmann at Cornell, these references weaken any potential protection OpenAI would possibly mount claiming the scenario is all an enormous coincidence. “They deliberately invited the general public to make the identification between Sky and Samantha. That is not a superb look,” Grimmelmann says. “I wonder if a lawyer reviewed Altman’s ‘her’ tweet.” Mixed with Johansson’s revelations that the corporate had certainly tried to get her to offer a voice for its chatbots—twice over—OpenAI’s insistence that Sky will not be meant to resemble Samantha is difficult for some to consider.
“It was a boneheaded transfer,” says David Herlihy, a copyright lawyer and music trade professor at Northeastern College. “A miscalculation.”
Different legal professionals see OpenAI’s conduct as so manifestly goofy they believe the entire scandal is perhaps a deliberate stunt—that OpenAI judged that it may set off controversy by going ahead with a sound-alike after Johansson declined to take part however that the eye it could obtain from appeared to outweigh any penalties. “What’s the purpose? I say it’s publicity,” says Purvi Patel Albers, a associate on the legislation agency Haynes Boone who usually takes mental property instances. “The one compelling motive—perhaps I’m giving them an excessive amount of credit score—is that everybody’s speaking about them now, aren’t they?”
